Transcript Lec17_heritability

Announcements
1. Make up exam:
Friday Nov 9, 2-4pm, Room BV363
(see blackboard for details)
2. Video 2 is posted on weboption:
‘Evolution: The Eternal Arms Race’
3.Office hours this week
TODAY SW551 11am to noon.
Friday SW551 & Virtual. 1pm – 2pm
4. Slide correction: Cystic Fibrosis example on p.214-218
5. End of HIV lecture will be provided as an audio file linked
to powerpoint by Monday.
Tropical
Biodiversity
Field Course
(BIOC51 Spring)
Costa Rica
The COURSE :
• Semester-long course with lectures at UTSC and a
8 day field trip to Costa Rica (during Feb break)
• Focus on biodiversity of tropical freshwater
ecosystems
• Learn field sampling techniques
• Participate in biological survey of the Costa Rica
fish fauna
• See amazing animals, plants, and habitats
The DETAILS:
•
•
•
•
Lectures: Spring, Tuesdays 11am-noon
Field trip: Feb 16th to Feb 24th
Cost: $1800 + tuition for 0.5 FCE
Contact N. Lovejoy or M. Kolmann as soon as
possible to reserve a spot (deadline November
15th)
• For more information, contact Prof Nathan
Lovejoy, [email protected]
• Or TA Matt Kolmann,
[email protected]
Lec 17: Quantitative Genetics & Heritability
• Quantitative traits
• Heritability & Evolution
Recall: from Population Genetics:
Simplification: Assume phenotypes fall into discrete
categories, determined strictly by genotypes
e.g.,
RR
e.g.,
Rr
GG, Gg
rr
gg
We can now assign some relative
lifetime reproductive success to each genotype
& use Hardy-Weinberg analyses
BUT: internal & external environment important
for development of most traits
genes
protein
genes
•e.g.,
•amino acids
•enzymes
Biotic & abiotic
factors
protein
•e.g.,
•other proteins
•ph
phenotype
Quantitative Genetics
Quantitative trait = a characteristic for
which phenotypes show continuous
variation among individuals (e.g., height)
• traits determined by the environment and
many mendelian genes (at many loci)
with small effects
Most traits
focus on:
• phenotypes & heritability
• statistical properties of population traits
(e.g., mean, variance)
e.g., Quantitative Traits
Colour
Stem length
Intelligence
Sprint speed
Plant form (height,
#stems) varies with:
1. High elevation
1. Environment
(elevation)
2. Medium elevation
AND
2. Genotype
height
(cm)
3. Low elevation
clones
e.g., Yarrow (Achillea)
Take plants with exactly the
same genotypes, rear in three
different environments
Genotypes
Plant form (height,
#stems) varies with:
1. High elevation
1. Environment
(elevation)
2. Medium elevation
AND
2. Genotype
height
(cm)
3. Low elevation
clones
e.g., Yarrow (Achillea)
Take plants with exactly the
same genotypes, rear in three
different environments
Genotypes
e.g., Quantitative Traits
Typical distribution:
normal distribution
(bell curve)
#
students
20
18
16
14
12
10
8
6
4
2
0
5.0 5.2 5.4 5.6 5.8 5.10 6.0 6.2 6.4
Height (feet.inches)
6.6
Problem: we usually do not know which loci
control quantitative traits
How do we study evolution of these traits?
Quantitative Genetics
Tools for measuring:
•Heritable variation in traits (heritability)
•Differences in lifetime reproductive success
(intensity of selection)
Tools for predicting:
•Effect of selection on phenotypic trait value
(Evolutionary response to selection)
Common mistake ALERT !!
This is NOT your
grandma’s
‘heritability’
‘HERITABLE’ has a
specific meaning
in evolutionary
biology and
has little to do with
the common usage of
the words ‘heritable’
or ‘inherited’
Heritability
Heritability? = Is this student 5’ tall due to genes
or environment?
•Due to her genes
•Due to the environment
Always
true
•Due to combined effect of genes and environment
Heritability
Heritability: Why is the shortest student shorter
than the tallest student?
•Differences in their genes
•Differences in their environments
•Differences in genes and environment
Heritability = extent to which phenotypic differences are due to differences in
genes underlying those traits
Heritability
Population level:
What fraction of the variation in height is due to
variation in genes (& what fraction is due to
variation in environment)?
‘broad sense heritability’
= fraction of total variation in
trait that is due to variation in genes
Heritability
Is the trait heritable?
Loosely: “Do offspring tend to resemble their
parents?”
(when we control for variation due to environment)
Why do we care about heritability?
The greater the relative effect of GENES on a trait, the
stronger the evolutionary response to selection
Need to know heritability to predict evolutionary
effects of selection
selection
parents
size
offspring
IF strong
genetic effect
on trait
High
heritability
Evolutionary change…
offspring
NO Evolutionary change…
IF strong
environmental
effect on trait
Low
heritability
Heritability
Population level:
What fraction of the variation in height is
due to variation in genes, what fraction is
due to variation in environment?
VP = phenotypic variation
= total variation in trait
VP = VG + VE
VG = genetic variation
= variation among individuals due to
variation in their genes
VE = environmental variation
= variation among individuals due to
variation in their environment
Different types of genetic variation
VP = VG + VE
VG = VA + VD
Dominance genetic variation
Differences among individuals
due to interactive effects of
genes
• e.g.,dominance, epistasis
Additive genetic variation
Differences among individuals due to additive
effects of genes
• contribution of each allele to phenotype is
independent of other alleles
Heritability
depends on
additive
genetic
variation
underlying
variation in
traits
Heritability depends on additive genetic
variation underlying variation in traits
Loosely:
additive genetic variation is the type of genetic
variation that leads to offspring looking like ~
‘average of their parents’
Defined to allow us to predict effect of selection
on phenotypic distributions
Different types of genetic variation
• simplified (qualitative) consideration of problem:
• two alleles, R and r
1. VA Additive genetic variance (no dominance)
Each ‘R’ allele
contributes one ‘unit’
of Red, each ‘r’ allele
one ‘unit’ of white
phenotype
rr
Rr
RR
genotype
every allele contributed by parent changes the phenotype
of offspring by one ‘unit’ (additive, predictable response to selection)
Different types of genetic variation
• simplified consideration of problem:
• two alleles, R and r
1. VA Additive genetic variance (no dominance)
Imagine:
phenotype
rr
Rr
RR
Strong
selection
for red
genotype
every allele contributed by parent changes the phenotype
of offspring by one ‘unit’ (additive, predictable response to selection)
Different types of genetic variation
• simplified consideration of problem:
• two alleles, R and r
1. VA Additive genetic variance (no dominance)
phenotype
Parents:
RR x RR
Next
generation:
rr
Rr
RR
All red
genotype
every allele contributed by parent changes the phenotype
of offspring by one ‘unit’ (additive, predictable response to selection)
Different types of genetic variation
Compare to:
phenotype
2. VD Dominance genetic variance
‘R’ allele is dominant,
‘r’ allele is recessive
rr
Rr
RR
effect of alleles contributed by parent on phenotype of
offspring depends on other alleles (response to selection not predictable)
Different types of genetic variation
Compare to:
2. VD Dominance genetic variance
phenotype
Imagine:
rr
Rr
RR
Strong
selection
for red
effect of alleles contributed by parent on phenotype of
offspring depends on other alleles (response to selection not predictable)
Different types of genetic variation
2. VD Dominance genetic variance
Parents:
phenotype
RR & Rr
Next
generation:
rr
Rr
RR
Red & White
effect of alleles contributed by parent on phenotype of
offspring depends on other alleles (response to selection not predictable)
Heritability
Narrow-sense heritability ( h2)
h2 =
VA =
VP
VA
VA + VD + VE
= fraction of total variation in trait
that is due to additive genetic variation
= h2 (0 < h2 < 1)
All environment
All genes
This is what we mean when we say
‘heritability’ in evolutionary studies
Measuring Heritability
A. Parent-offspring regression:
how much do offspring resemble their parents?
(when we control for variation due to environment)
Measure of the amount of
variation in parental trait due to
additive genetic variation
Measuring Heritability
A. Parent-offspring regression:
how much do offspring resemble their parents?
Slope = 0.84
Mid-offspring
height (inches)
rise
run
Mid-parent height (inches)
h2 = slope of parent-offspring regression
Measuring Heritability
A. Parent-offspring regression:
how much do offspring resemble their parents?
Heritability  0.5
Heritability  0
Heritability  1.0
10
Mid-offspring
0
trait
0
10
0
10
Mid-parent trait
Most traits
0
10
Measuring Heritability
A. Parent-offspring regression:
BUT: Parents may resemble offspring due to:
•similar genotype
•similar environment
To measure heritability accurately, must control
for environmental effects
Measuring Heritability
Problem: similarities may be due to similar environments...
Good diet
Mid-offspring
height (inches)
h2 = 0.84
?
Poor diet
Mid-parent height (inches)
Poor diet
Good diet
Measuring Heritability
A. Parent-offspring regression:
Problem: similarities may be due to similar environments...
Control for environment = isolate genetic effects
Strategies
1) Ensure offspring and parents have
different environments
e.g., cross-fostering / adoption (animals) or
reciprocal transplant (plants) experiments
2) Ensure all offspring have same
environment
e.g., ‘common garden’ experiments
Measuring Heritability
A. Parent-offspring regression:
1. Cross-fostering experiments
Song sparrow
Heritability of beak depth
Measuring Heritability
A. Parent-offspring regression:
1. Cross-fostering experiments
Offspring are reared
by non-genetic
parents
(foster parents)
Measuring Heritability
A. Parent-offspring regression:
1. Cross-fostering experiments
Measure beak depths of adult offspring,
genetic parents & foster parents
Measuring Heritability
A. Parent-offspring regression (1. cross fostering)
(different environment, any similarity due to genes)
Accurate estimate of h2
Reared in nest of ‘foster parent’
Measuring Heritability
A. Parent-offspring regression:
Genetic parents & cross-fostered
offsprings
Foster parents & foster-offspring
(same environment, different genes)
Compare
to 
Similarity is due to underlying genetic similarity rather
than environmentally-induced similarities
Measuring Heritability
A. Parent-offspring regression:
2. Common-garden experiment
Common Lab
Environment
Measuring heritability
B) Twin studies
(i) monozygotic
twins reared
apart
•identical genotype
•different environment
how similar
are they as
adults?
Jack Yufe
raised Jewish in
Carribbean
Oskar Stohr
raised Catholic in
Nazi Germany
Measuring heritability
Jack Yufe
Oskar Stohr
despite 47 years apart they both
“like sweet liqueurs, …store rubber bands on their wrists,
read magazines back to front, dip buttered toast in their
coffee and have highly similar personalities”
Holden . 1980. Science. 207:1323-1328
Measuring heritability
B) Twin studies
(ii) Compare
similarity of
monozygotic
twins to
similarity of
dizygotic twins
Clones
Siblings
100% genetic
similarity
50% genetic
similarity
monozygotic
dizygotic
Monozygotic:
•Same genes
•Same environment
monozygotic
Dizygotic:
•different genes
•same environment
Heritability high
>
Heritability low
=
dizygotic
Heritability: a common error
Heritability is NOT A FIXED
characteristic of a trait
Heritability is specific to a particular
population in a particular environment
Why??
Heritability
h2 =
VA =
VP
VA
VG +VE
What happens to h2 if VE changes?
h2 will change
•Even if VA underlying trait remains the SAME
Heritability is not a fixed value: hypothetical example
e.g., What is the heritability of skin colour among
Caucasians in Vancouver?
Heritability is not a fixed value: hypothetical example
e.g., What is the heritability of skin colour among
Caucasians in Vancouver?
Facts: sun exposure is the largest environmental determinant of
skin colour
• Vancouver: winter (hardly any sun, rains all the time); summer
is quite sunny
Assume: random variation
in how much time people
spend outdoors (tanning)
when it’s sunny (VE)
Heritability = VA / (VG + VE)
Winter
VE is low
Prediction: h2 is high
dark
Mid
-offspring
skin tone
light
light
dark
Mid-parent skin tone
Summer
VE is high
Prediction: h2 is low
dark
Mid
-offspring
skin tone
light
light
dark
Mid-parent skin tone
Heritability: common errors 1
Error 1. If all members of a population have
the same trait value, that trait is highly
heritable
What is the
heritability of
having a nose in
humans?
Heritability: common errors 1
Error 1. If all members of a population have
the same trait value, that trait is highly
heritable
Correction:
If there is no variation in a trait, the trait is
NOT heritable
Heritability = VA / VP
= 0/0
Heritability: common errors 2
Error 2. Heritability is a fixed value for a
particular trait
Correction:
Heritability if NOT a fixed characteristic of a
trait
Heritability of a trait is specific to a
particular population in a particular
environment
Heritability: common errors 3
Error 3. ‘Heritability tells us whether
differences between populations are due to
differences in genes or the environment’
Heritability
Does heritability tell us something about
the source of differences between populations?
Hypothetical example: you observe that people
from Vancouver have fairer skin than people from
Los Angeles on average
Vancouver
Los Angeles
Light
Dark
You also find: High heritability for skin
tone within each population
Winter, Vancouver
Winter, Los Angeles
h2 high
Mid
-offspring
h2 high
Mid
-offspring
Mid-parent
Mid-parent
Does this mean that the difference in average skin colour of
people from Vancouver versus California is due to
genetic differences between the populations?
Does this mean that the difference in average skin colour of
people from Vancouver versus California is due to
genetic differences between the populations?
Jan – March in
Los Angeles
Jan – March in
Vancouver
Is this difference in skin colour due to a genetic difference?
Does this mean that the difference in average skin colour of
people from Vancouver versus California is due to
genetic differences between the populations?
NO
One likely hypothesis:
It is sunnier in Los Angeles than in Vancouver, thus Los
Angeleans are darker skinned on average
e.g., in this case, difference between populations
mainly due to VE despite high heritability within each
population
Heritability: common errors 3
Error 3. ‘Heritability tells us whether
differences between populations are due to
differences in genes or the environment’
CORRECTION:
Heritability does not tell you anything
about the cause of differences BETWEEN
populations
e.g., Yarrow
(Achillea)
1. High elevation
High heritability
WITHIN each
population
reminder
2. Medium elevation
height
(cm)
Why is there
variation in plant
form between
populations?
3. Low elevation
Genotypes
SO:
•Heritability is NOT a fixed characteristic of a trait
(varies with environment & across populations)
•Is NOT useful for identifying the source of
differences BETWEEN populations
Why are you messing with my mind????
Did you just waste this entire lecture on nonsense???
SO:
•Heritability is NOT a fixed characteristic of a trait
(varies with environment & across populations)
•Is NOT useful for identifying the source of
differences BETWEEN populations
Why bother to measure heritability?
Allows prediction of whether selection on
a trait in a given population will cause that
trait to evolve
Measuring the response to selection
Evolutionary response to selection
Magnitude & direction of change in trait in
offspring
Depends on:
1. Heritability
•
R=
2
hS
extent to which parental
phenotype predicts
offspring phenotype
2. Selection differential
•magnitude & direction of selection
Measuring Heritability
Domestic cattle
Trait
Gestation period
Milk yield
Height
h2
0.3
0.3
0.6
Measuring Heritability
Domestic Trait
Litter
size
pigs
Back fat thickness
h2
0.1
0.64
NEXT LECTURES
The (abuse and) Misuse of Heritability